1. Volume 18, Issue 6, Pages 687-698 (June 2005)
Structural Determinants for Selective Recognition of a Lys48-Linked Polyubiquitin Chain by a UBA Domain 
Ranjani Varadan, Michael Assfalg, Shahri Raasi, Cecile Pickart, David Fushman 
Molecular Cell 
Volume 18, Issue 6, Pages (June 2005)
DOI: /j.molcel
Copyright © 2005 Elsevier Inc. Terms and Conditions

2. Figure 1 CSP Mapping of the Sites in Ub2 Involved in UBA2 Binding
(A and B) CSPs, Δδ, are shown in the (A) distal and (B) proximal Ub at saturation ([UBA2]:[Ub2] = 5). Asterisks denote residues showing signal attenuation characteristic of intermediate exchange. Gray bars in (B) mark residues in slow exchange.
(C) Surface representation of Ub2, with the distal and proximal domains colored marine and light blue, respectively, and the perturbed sites painted red (distal Ub) and yellow (proximal). The two domains are also shown separated to highlight the location of perturbed sites at the Ub:Ub interface.
(D) Distal and proximal Ubs respond differently during titration of UBA2. Shown are regions of the 15N-1H HSQC spectra corresponding to Thr14 for the distal (top row) and proximal (bottom row) Ub in Ub2 at the [UBA2]:[Ub2] molar ratios indicated on the top. The behavior of Thr14 signals in the proximal Ub corresponds to a slow exchange regime.
(E) Representative binding curves for the indicated residues in the distal Ub. The Kd values obtained from the fit (see the Experimental Procedures) were 237, 74, and 170 μM for the distal Ub and 6, 5, and 2 μM for the proximal Ub (based on data for Ile13, Phe45, and Leu69, respectively).
The drawings of protein structures in this paper were generated using MolMol (Koradi et al., 1996) and PyMol (DeLano, 2002).
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

3. Figure 2
CSP Mapping of the UBA2 Epitope Involved in Binding monoUb and Lys48-Linked Ub Chains
CSPs in UBA2 upon addition of saturating amounts of (A) monoUb, (B) Ub2, and (C) Ub4. On the right are surface maps of the perturbed residues (magenta) in the presence of (D) monoUb and (E) Ub2, and a ribbon representation (F) of the UBA2 structure colored according to (E). In (C), residues 342–345 (triangles) in UBA-helix 2 and loop L2 could not be reliably detected due to signal broadening (intermediate exchange) upon addition of Ub4. UBA2 binding to Lys63-linked Ub2 results in a CSP pattern similar to that in (A) (Varadan et al., 2004).
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

4. Figure 3
Molecular Mass of UBA2:Ub2 Complexes Determined by 15N T1 Measurements
The calibration curve was obtained assuming isotropic overall tumbling of the proteins in solution. The NH backbone dynamics were characterized by the squared order parameter of 0.83 and local correlation time of 10 ps. We assumed a protein specific density of 0.76 cm3/g and included a hydration shell that causes about 50% increase in the overall tumbling time. Vertical dashed lines indicate the expected molecular mass values for Ub, Ub2, and Ub4 complexes with UBA2, as indicated. The horizontal lines (with error bars on the right) represent the average values of 15N T1 measured for Lys48-linked Ub2:UBA complexes at the indicated molar ratios. Data for complexes of UBA2 with monoUb and Lys63-linked Ub2 are also shown. The circles depict experimental data for monoUb, Ub2 (Varadan et al., 2002), Ub4, and GB3 (Hall and Fushman, 2003) used here to build the calibration curve. The error bars represent standard deviations over multiple residues in a given protein. The T1 data for Ub4 are from our measurements and also converted from those reported in Tenno et al. (2004) for a different temperature and spectrometer frequency.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

5. Figure 4
A Subset of Intermolecular NOE Signals Demonstrating Direct Interactions between UBA2 and Lys48-Linked Ub2
Shown are representative regions of the 15N-edited NOESY spectra (200 ms mixing time) corresponding to NOE signals originating on the methyl hydrogens in Ub2 and detected for the backbone amides in 15N-2H-UBA2. The data correspond to (A) fully protonated Ub2, (B) only proximal Ub protonated (distal Ub perdeuterated), and (C) only distal Ub protonated. UBA2 residues and the helices they belong to are indicated. A slight horizontal shift in the peak positions between the spectra reflects variations in the Ub2:UBA2 molar ratio between the samples. These shifted peak positions were verified by comparison with the 1H-15N HSQC maps recorded for the same samples.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

6. Figure 5 Docked Model of the Structure of the 1:1 UBA2:Ub2 Complex
The structure was obtained with a combination of CSPs (ambiguous restraints) and NOEs (unambiguous restraints) as described in the Experimental Procedures. (A) Ribbon diagram of the complex; Ub domains are in green, UBA2 is in wheat. Side chains of several residues involved in NOE contacts and in the Lys48-Gly76 linkage (cyan) are shown in stick. Also shown are UBA2-Arg326 and Leu330 making contact with Gly47 and Lys48 of the distal Ub. The isopeptide bond is indicated by a dashed line. (B and C) Surface representations of the Ub2 part of the complex with the binding pocket residues painted purple. The UBA domain is shown as a ribbon. The orientation in (C) is rotated by 60° around the (horizontal) interdomain axis relative to (A) and (B). (D) Experimental verification of the UBA2:Ub2 complex structure by site-specific paramagnetic spin labeling. Colored blue are sites that showed relaxation rate enhancement (signal attenuation) when a spin label, 1-oxyl-2,2,5,5-tetramethyl-3-pyrroline-3-methyl)methanesulfonate, was attached to the sole cysteine of UBA2 (Cys344). UBA2 residues that showed signal attenuation when SL was attached to Cys48 of the distal Ub are colored red. The location of the SL attachment sites in these two experiments is indicated by the blue and red spheres, respectively. No significant changes in peak positions in the HSQC spectra occurred as a result of spin labeling of the binding partner, indicating no interference of the SL with the UBA2-Ub2 interaction.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

7. Figure 6
Recognition of Lys48-Linked Ub2 Involves a Unique Epitope on UBA2
A comparison of the orientation of the UBA2 domain on the surface of the distal (A) and proximal (B) Ub units in Ub2 with the published structures of monoUb complexes with (C) UBA2 and (D) UBA1 domains from hHR23A (Mueller et al., 2004) and (E) CUE domain from Cue2 (Kang et al., 2003). Ub is shown in green, and the UBA (or CUE) domains are in wheat.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions